An Observational Study of Honey Bee Colony Winter Losses and Their Association with Varroa destructor, Neonicotinoids and Other Risk Factors
These effects were estimated from a binomial GLMM allowing for significant fixed effects for beekeeper risk of colony loss in the Netherlands and also with a beekeeper random effect, for the 1 to 50 colony operations used in the model fitting. The data are derived from the Dutch National Monitoring Survey for Honey Bee Winter Loss 2012. The legend within the figure shows the key to the colour coding. Darker green indicates areas of lower risk and darker red areas of higher risk of loss. The value of the postal code effect was used as a predictor in the final model of the data in the present study.